Enhancing Organizational Knowledge with Visual Documentation: Implementing Amazon Q Business and Custom Document Enrichment

Bridging the Gap: Transforming Visual Data into Actionable Insights

Use Case: Unlocking Educational Demographics Through Visual Analytics

Step-by-Step Guide: Implementing Custom Document Enrichment in Amazon Q Business

Solution Overview: Integrating Image Processing for Enhanced Searchability

Prerequisites for Implementation: Setting Up Your Amazon Q Business Environment

Creating and Syncing an Amazon Q Business Application with S3

Optimizing Performance: Configuring CDE for Amazon S3 Data Sources

Extracting Insights from Images: Leveraging Amazon Bedrock’s Capabilities

Example Prompts and Results: Interacting with Visual Data

Best Practices for CDE Configuration in Amazon S3

Conclusion: Unlocking the Potential of Visual Data with AWS Services

About the Authors: Meet the Experts Behind the Integration

Unlocking the Power of Visual Data with Amazon Q Business and Custom Document Enrichment

In today’s digital landscape, conveying complex information effectively is paramount. Organizations often rely on visual documentation—like diagrams, charts, and technical illustrations—to communicate intricate ideas. However, despite the integration of text documents in modern knowledge management systems, much of the rich information contained in visuals frequently remains inaccessible to search and AI systems. This gap can lead to significant misunderstandings and hinder decision-making processes.

Enter Amazon Q Business and its Custom Document Enrichment (CDE) feature, which expands the capabilities of visual data processing. By enabling organizations to analyze and reinterpret standalone image files, such as JPGs and PNGs, businesses can extract crucial insights from previously dormant visual information. This blog post guides you through implementing the CDE feature within an Amazon Q Business application, showcasing its impact on organizational knowledge bases.

Example Scenario: Analyzing Regional Educational Demographics

Imagine you’re part of a national educational consultancy that stores charts and graphs portraying demographic data across various AWS regions in an Amazon S3 bucket. A particular image illustrates student distribution by age range across different cities using a bar chart. Such visualizations are gold mines for decision-making but often remain locked in image formats in S3 buckets.

With Amazon Q Business and the CDE feature, you can enable natural language queries against these visual data points. For instance, your team might ask, “Which city has the highest number of students in the 13–15 age range?” or “Compare the student demographics between City 1 and City 4” directly through the Amazon Q Business interface.

Bridging the Gap with Amazon Q Business

Utilizing the CDE feature, organizations can:

Detect and process image files during document ingestion.
Leverage Amazon Bedrock with AWS Lambda to interpret visual information.
Extract structured data and insights from charts and graphs.
Make this information easily searchable via natural language queries.

Solution Overview

In this solution, we will outline how to implement a CDE-based solution for your educational demographic data visualizations. Amazon Q Business automates the extraction of meaning from image files during the ingestion process. The CDE rules trigger an AWS Lambda function upon encountering the S3 path, identifying image files and calling the Amazon Bedrock API to analyze and extract contextual information.

When processed, the extracted text will be integrated into your knowledge base in Amazon Q Business. Users can then search for impactful data and insights founded on actual context, unlocking insights previously hidden within their image repositories.

Implementation Steps

Here’s a high-level overview of the steps you’ll need to take:

Create an Amazon Q Business Application: Sync it with your S3 bucket.
Configure the CDE feature: Set it up within the Amazon Q Business application for your S3 data source.
Extract context from the images: Enable AWS Lambda to facilitate this extraction.

Prerequisites

Before diving in, ensure you have:

An AWS account.
At least one Amazon Q Business Pro user with admin permissions.
IAM permissions to create/manage roles and policies.
A compatible data source, like an Amazon S3 bucket.
Access to an Amazon Bedrock LLM in the appropriate AWS Region.

Detailed Implementation Steps

Step 1: Creating an Amazon Q Business Application

Start by either using the AWS Management Console or CLI to create your Amazon Q Business application. Create an index and utilize the built-in Amazon S3 connector to link your application with documents stored in your organization’s S3 bucket.

Step 2: Configuring CDE for Amazon S3

With the CDE feature, you’ll apply enhancements, filtering, and modifications that augment enterprise content during the ingestion process. To configure the CDE for your Amazon S3 data source:

Select your application and navigate to Data sources.
Choose your existing S3 data source or create a new one, ensuring that multimedia content settings are correctly configured.
Locate the Custom Document Enrichment section and set the pre-extraction rules to activate an AWS Lambda function when specific S3 conditions are met.

Step 3: Extracting Context from Images

To extract insights, the Lambda function calls the Amazon Bedrock API using a specific model. The code samples below encapsulate the Lambda function design:

import boto3
import logging
import json
from typing import List, Dict, Any
from botocore.config import Config

MODEL_ID = "us.anthropic.claude-3-7-sonnet-20250219-v1:0"
MAX_TOKENS = 2000
MAX_RETRIES = 2
FILE_FORMATS = ("jpg", "jpeg", "png")

logger = logging.getLogger()
logger.setLevel(logging.INFO)
s3 = boto3.client('s3')
bedrock = boto3.client('bedrock-runtime', config=Config(read_timeout=3600, region_name="us-east-1"))

def lambda_handler(event, context):
    ...

Prompting the Model

Constructing the prompts is critical for obtaining the desired output from your images. Tailor the structure based on your specific use case by making use of various model features available through Amazon Bedrock.

Example Prompts and Results

For instance, after processing a student age distribution graph, your queries could yield results like:

Q: Which City has the highest number of students in the 13-15 age range?
A: City X has the highest count with Y students.

Such automated interaction with previously static data transforms the way users engage with visual content.

Best Practices for CDE Configuration

Utilize conditional rules to only process file types needing transformation.
Monitor Lambda execution through Amazon CloudWatch to capture errors and performance metrics.
Set adequate timeout values for optimal Lambda function execution.

Cleanup

To ensure you’re not incurring unnecessary charges, follow these steps to clean up:

Remove subscriptions for users/groups within Amazon Q Business.
Delete the Amazon Q Business application.
Remove the Lambda function.
Empty and delete the S3 bucket.

Conclusion

Combining Amazon Q Business, Custom Document Enrichment, and Amazon Bedrock can revolutionize the approach toward visual data. This transformation allows organizations to bridge the gap between visual content and actionable insights, enabling users to query complex imagery simply and effectively.

Explore Amazon Q Business and Amazon Bedrock documentation for further insights and specific use case implementations to unlock the potential of your visual data today!

Exclusive Content:

Extracting Context from Image Files in Amazon Q Business with LLMs